In the Internet age, people are becoming more and more familiar in experiencing online services. In many cases, the customer commits herself and her assets in a business transaction with no (or limited) possibility to test the service/good she is booking/buying. Hence, there is the need to prove the trustworthiness of such services for supporting a user in her choice. Many websites feed the customer with reviews of past users representing their degree of satisfaction. In this paper, we consider a scenario where different services may be grouped together to form packets, and we design and implement a simple procedure through which a customer can choose the packet that best satisfies her expectations. The final choice will be driven both by the qualities of the reviews on the constituting services, and by the customer's personal preference and attitudes. To automatise the procedure, we survey real behaviours of users when they choose a service and give reviews, by obtaining a probabilistic model plugged in our simulator. In particular, we deal with the issue of false review, reported by unfair users that intentionally act malevolently. The simulations results show that our system is robust enough up to a certain number of unfair feedback.
A study on rating services based on users' categories
Gianpiero Costantino;Fabio Martinelli;Marinella Petrocchi
2012
Abstract
In the Internet age, people are becoming more and more familiar in experiencing online services. In many cases, the customer commits herself and her assets in a business transaction with no (or limited) possibility to test the service/good she is booking/buying. Hence, there is the need to prove the trustworthiness of such services for supporting a user in her choice. Many websites feed the customer with reviews of past users representing their degree of satisfaction. In this paper, we consider a scenario where different services may be grouped together to form packets, and we design and implement a simple procedure through which a customer can choose the packet that best satisfies her expectations. The final choice will be driven both by the qualities of the reviews on the constituting services, and by the customer's personal preference and attitudes. To automatise the procedure, we survey real behaviours of users when they choose a service and give reviews, by obtaining a probabilistic model plugged in our simulator. In particular, we deal with the issue of false review, reported by unfair users that intentionally act malevolently. The simulations results show that our system is robust enough up to a certain number of unfair feedback.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
